Hemophilia A is the most common hereditary coagulation disorder, with an estimated incidence of 1 per 5000 males. It is caused by deficiency or structural defects in FVIII, a critical component of the intrinsic pathway of blood coagulation. The current treatment for hemophilia A involves intravenous injection of human FVIII. Human FVIII has been produced recombinantly as a single-chain molecule of approximately 300 kD. It consists of the structural domains A1-A2-B-A3-C1-C2 (Thompson, 2003, Semin. Hematol. 29, pp. 11-22). The precursor product is processed into two polypeptide chains of 200 kD (heavy) and 80 kD (light) in the Golgi Apparatus, with the two chains held together by metal ions (Kaufman et al., 1988, J. Biol. Chem. 263, p. 6352; Andersson et al., 1986, Proc. Natl. Acad. Sci. 83, p. 2979).
The B-domain of FVIII seems to be dispensable as B-domain deleted FVIII (BDD, 90 kD A1-A2 heavy chain plus 80 kD light chain) has also been shown to be effective as a replacement therapy for hemophilia A. The B-domain deleted FVIII sequence contains a deletion of all but 14 amino acids of the B-domain.
Hemophilia A patients are currently treated by intravenous administration of FVIII on demand or as a prophylactic therapy administered several times a week. For prophylactic treatment 15-25 IU/kg bodyweight is given of factor VIII three times a week. It is constantly required in the patient. Because of its short half-life in man, FVIII must be administered frequently. Despite its large size of greater than 300 kD for the full-length protein, FVIII has a half-life in humans of only about 11 hours. (Ewenstein et al, 2004, Semin. Hematol. 41, pp. 1-16). The need for frequent intravenous injection creates tremendous barriers to patient compliance. It would be more convenient for the patients if a FVIII product could be developed that had a longer half-life and therefore required less frequent administration. Furthermore, the cost of treatment could be reduced if the half-life were increased because fewer dosages may then be required.
An additional disadvantage to the current therapy is that about 25-30% of patients develop antibodies that inhibit FVIII activity (Saenko et al, 2002, Haemophilia 8, pp. 1-11). The major epitopes of inhibitory antibodies are located within the A2 domain at residues 484-508, the A3 domain at residues 1811-1818, and the C2 domain. Antibody development prevents the use of FVIII as a replacement therapy, forcing this group of patients to seek an even more expensive treatment with high-dose recombinant Factor VIIa and immune tolerance therapy.
The following studies identified FVIII epitopes of inhibitory antibodies. In a study of 25 inhibitory plasma samples, 11 were found to bind to the thrombin generated 73 kD light chain fragment A3C1C2, 4 to the A2 domain, and 10 to both (Fulcher, C. et al., 1985, Proc. Natl. Acad. Sci. 2(22), pp. 7728-32). In another study, six of eight A2 domain inhibitors from patients were neutralized by a recombinant A2 polypeptide (Scandella, D. et al., 1993, Blood 82(6), pp. 1767-75). Epitopes for six of nine inhibitors from patients were mapped to A2 residues 379-538 (Scandella, D. et al., 1988, Proc. Natl. Acad. Sci. 85(16), pp. 6152-6). An epitope for 18 heavy-chain inhibitors was localized to the same N-terminal 18.3 kD region of the A2 domain (Scandella, D. et al., 1989, Blood 74(5), pp. 1618-26).
An active, recombinant hybrid human/porcine FVIII molecule, generated by replacing human A2 domain residues 387-604 with the homologous porcine sequence, was resistant to a patient A2 inhibitor (Lubin, I. et al., 1994, J. Biol. Chem. 269(12), pp. 8639-41) and resistant to a murine monoclonal antibody mAB 413 IgG that competes with patient A2 inhibitors for binding to A2 (Scandella, D. et al., 1992, Thromb Haemost. 67(6), pp. 665-71). This A2 domain epitope was further localized to the A2 domain residues 484-508 when experiments showed that mAB 413 IgG and four patient inhibitors did not inhibit a hybrid human/porcine FVIII in which the A2 domain residues 484-508 were replaced with that of porcine (Healey, J. et al., 1995, J. Biol. Chem. 270(24), pp. 14505-9). This hybrid FVIII was also more resistant to at least half of 23 patient plasmas screened (Barrow, R. et al., 2000, Blood 95(2), pp. 564-8). Alanine scanning mutagenesis identified residue 487 to be critical for binding to all five patient inhibitors tested, while residues 484, 487, 489, and 492 were all important to interaction with mAB 413 IgG (Lubin, I., J. Biol. Chem. 272(48), pp. 30191-5). Inhibitory antibody titers in mice receiving the R484A/R489A/P492A mutant, but not the R484A/R489A mutant, were significantly lower than in mice receiving control human BDD FVIII (Parker, E. et al., 2004, Blood 104(3), pp. 704-10). In sum, the 484-508 region of the A2 domain seems to be a binding site for inhibitors of FVIII activity.
In addition to the development of an immune response to FVIII, another problem with conventional therapy is that it requires frequent dosaging because of the short half-life of FVIII in vivo. The mechanisms for clearance of FVIII from the circulation have been studied.
FVIII clearance from circulation has been partly attributed to specific binding to the low-density lipoprotein receptor-related protein (LRP), a hepatic clearance receptor with broad ligand specificity (Oldenburg et al., 2004, Haemophilia 10 Suppl 4, pp. 133-139). Recently, the low-density lipoprotein (LDL) receptor was also shown to play a role in FVIII clearance, such as by cooperating with LRP in regulating plasma levels of FVIII (Bovenschen et al., 2005, Blood 106, pp. 906-910). Both interactions are facilitated by binding to cell-surface heparin sulphate proteoglycans (HSPGs). Plasma half-life in mice can be prolonged by 3.3-fold when LRP is blocked or 5.5-fold when both LRP and cell-surface HSPGs are blocked (Sarafanov et al., 2001, J. Biol. Chem. 276, pp. 11970-11979). HSPGs are hypothesized to concentrate FVIII on the cell surface and to present it to LRP. LRP binding sites on FVIII have been localized to A2 residues 484-509 (Saenko et al., 1999, J. Biol. Chem. 274, pp. 37685-37692), A3 residues 1811-1818 (Bovenschen et al., 2003, J. Biol. Chem. 278, pp. 9370-9377) and an epitope in the C2 domain (Lenting et al., 1999, J. Biol. Chem. 274, pp. 23734-23739).
FVIII is also cleared from circulation by the action of proteases. To understand this effect, one must understand the mechanism by which FVIII is involved in blood coagulation. FVIII circulates as a heterodimer of heavy and light chains, bound to vWF. VWF binding involves FVIII residues 1649-1689 (Foster et al., 1988, J. Biol. Chem. 263, pp. 5230-5234), and parts of C1 (Jacquemin et al., 2000, Blood 96, pp. 958-965) and C2 domains (Spiegel, P. et al., 2004, J. Biol. Chem. 279(51), pp. 53691-8). FVIII is activated by thrombin, which cleaves peptide bonds after residues 372, 740, and 1689 to generate a heterotrimer of A1, A2, and A3-C1-C2 domains (Pittman et al., 2001, Proc. Natl. Acad. Sci. 276, pp. 12434-12439). Upon activation, FVIII dissociates from vWF and is concentrated to the cell surface of platelets by binding to phospholipid. Phospholipid binding involves FVIII residues 2199, 2200, 2251, and 2252 (Gilbert et al., 2002, J. Biol. Chem. 277, pp. 6374-6381). There it binds to FIX through interactions with FVIII residues 558-565 (Fay et al., 1994, J. Biol. Chem. 269, pp. 20522-20527) and 1811-1818 (Lenting et al., 1996, J. Biol. Chem. 271, pp. 1935-1940) and FX through interactions with FVIII residues 349-372 (Nogami et al., 2004, J. Biol. Chem. 279, pp. 15763-15771) and acts as a cofactor for FIX activation of FX, an essential component of the intrinsic coagulation pathway. Activated FVIII (FVIIIa) is partly inactivated by the protease activated protein C (APC) through cleavage after FVIII residues 336 and 562 (Regan et al., 1996, J. Biol. Chem. 271, pp. 3982-3987). The predominant determinant of inactivation, however, is the dissociation of the A2 domain from A1 and A3-C1-C2 (Fay et al., 1991, J. Biol. Chem. 266, pp. 8957-8962).
One method that has been demonstrated to increase the in vivo half-life of a protein is PEGylation. PEGylation is the covalent attachment of long-chained polyethylene glycol (PEG) molecules to a protein or other molecule. The PEG can be in a linear form or in branched form to produce different molecules with different features. Besides increasing the half-life of peptides or proteins, PEGylation has been used to reduce antibody development, protect the protein from protease digestion and keep the material out of the kidney filtrate (Harris et al., 2001, Clinical Pharmacokinetics 40, pp. 539-51). In addition, PEGylation may also increase the overall stability and solubility of the protein. Finally, the sustained plasma concentration of PEGylated proteins can reduce the extent of adverse side effects by reducing the trough to peak levels of a drug, thus eliminating the need to introduce super-physiological levels of protein at early time-points.
Random modification of FVIII by targeting primary amines (N-terminus and lysines) with large polymers such as PEG and dextran has been attempted with varying degree of success (WO94/15625, U.S. Pat. No. 4,970,300, U.S. Pat. No. 6,048,720). The most dramatic improvement, published in a 1994 patent application (WO94/15625), shows a 4-fold half-life improvement but at a cost of 2-fold activity loss after reacting full-length FVIII with 50-fold molar excess of PEG. WO2004/075923 discloses conjugates of FVIII and polyethylene glycol that are created through random modification. Randomly PEGylated proteins, such as interferon-alpha (Kozlowski et al, 2001, BioDrugs 15, pp. 419-429) have been approved as therapeutics in the past.
This random approach, however, is much more problematic for the heterodimeric FVIII. FVIII has hundreds of potential PEGylation sites, including the 158 lysines, the two N-termini, and multiple histidines, serines, threonines, and tyrosines, all of which could potentially be PEGylated with reagents primarily targeting primary amines. For example, the major positional isomer for PEGylated interferon Alpha-2b was shown to be a histidine (Wang et al., 2000, Biochemistry 39, pp. 10634-10640). Furthermore, heterogeneous processing of full length FVIII can lead to a mixture of starting material that leads to further complexity in the PEGylated products. An additional drawback to not controlling the site of PEGylation on FVIII is a potential activity reduction if the PEG were to be attached at or near critical active sites, especially if more than one PEG or a single large PEG is conjugated to FVIII. Because random PEGylation will invariably produce large amounts of multiply PEGylated products, purification to obtain only mono-PEGylated products will drastically lower overall yield. Finally, the enormous heterogeneity in product profile will make consistent synthesis and characterization of each lot nearly impossible. Since good manufacturing requires a consistent, well-characterized product, product heterogeneity is a barrier to commercialization. For all these reasons, a more specific method for PEGylating FVIII is desired.
Various site-directed protein PEGylation strategies have been summarized in a recent review (Kochendoerfer, G., Curr. Opin. Chem. Biol. 2005, available online as of Oct. 15, 2005, direct object identifier doi:10.1016/i.cbpa.2005.10.007). One approach involves incorporation of an unnatural amino acid into proteins by chemical synthesis or recombinant expression followed by the addition of a PEG derivative that will react specifically with the unnatural amino acid. For example, the unnatural amino acid may be one that contains a keto group not found in native proteins. However, chemical synthesis of proteins is not feasible for a protein as large as FVIII. Current limit of peptide synthesis is about 50 residues. Several peptides can be ligated to form a larger piece of polypeptide, but to produce even the B-domain deleted FVIII would require greater than 20 ligations, which would result in less than 1% recovery even under ideal reaction condition. Recombinant expression of proteins with unnatural amino acids has so far mainly been limited to non-mammalian expression systems. This approach is expected to be problematic for a large and complex protein such as FVIII that needs to be expressed in mammalian systems.
Another approach to site-specific PEGylation of proteins is by targeting N-terminal backbone amine with PEG-aldehydes. The low pH required under this process to achieve specificity over other amine groups, however, is not compatible with the narrow near-neutral pH range needed for the stability of FVIII (Wang et al., 2003, International J. Pharmaceutics 259, pp. 1-15). Moreover, N-terminal PEGylation of FVIII may not lead to improved plasma half-life if this region is not involved in plasma clearance. In fact, the N-terminal region of the FVIII light chain has been implicated in binding to the von Willebrand factor (vWF), a carrier protein that is critical for FVIII survival in circulation. By N-terminal modification of factor VIII, the critically important association with vWF may be disrupted or weakened. Thus, N-terminal PEGylation of FVIII may have the opposite effect of reducing plasma half-life of FVIII.
WO90/12874 discloses site-specific modification of human IL-3, granulocyte colony stimulating factor and erythropoietin polypeptides by inserting or substituting a cysteine for another amino acid, then adding a ligand that has a sulfhydryl reactive group. The ligand couples selectively to cysteine residues. Modification of FVIII or any variant thereof is not disclosed.
For the reasons stated above, there exists a need for an improved FVIII variant that possesses greater duration of action in vivo and reduced immunogenicity, while retaining functional activity. Furthermore, it is desirable that such a protein be produced as a homogeneous product in a consistent manner.